Plasma proteome correlations with liver stiffness in pediatric cholestasis implicate epithelial to mesenchymal transition.

Abstract

Background: Pediatric cholestatic liver diseases can be characterized by rapidly progressive fibrosis. A multicenter cross-sectional analysis of vibration-controlled elastography in biliary atresia (BA), alpha-1 antitrypsin deficiency (A1AT), and Alagille syndrome (ALGS) was leveraged to interrogate the plasma proteome relative to liver stiffness measurements (LSM).

Methods: Slow off-rate modified aptamer scanning profiling of >7000 proteins in plasma from 187 children with BA (n=93), A1AT (n=31), ALGS (n=46), and healthy pediatric controls (n=17) was performed, and correlations with LSM were undertaken.

Results: There was an abundance of LSM correlated proteins (BA n=2720, A1AT n=694, ALGS n=5968). Interestingly, a distinct plasma proteome was found in ALGS relative to BA and A1AT. Weighted Correlation Network Analysis identified groups of proteins with strong LSM correlation (eg, in a BA module of interest, Pearson correlation coefficient 0.79, p=5´0-21). Machine learning developed models predicting LSM as a continuous variable (median R2=0.62 for BA). For BA, time to transplant could be predicted equally well by the proteome or clinical parameters (elastic net models achieved a C-index using proteome 0.91, clinical parameters 0.91, proteome and clinical parameters 0.90). Single-cell transcriptomics predicted the potential hepatic cell of origin for the most informative proteins, which included macrophage, mesenchymal, mesothelial, and endothelial cells. The epithelial-to-mesenchymal transition pathway was enriched in LSM correlated proteins in all 3 diseases.

Conclusions: The plasma proteome is highly correlated in a disease-specific fashion with LSM in BA, A1AT, and ALGS. These correlations provide unique opportunities to identify biomarkers and focus attention on epithelial-to-mesenchymal transition in pediatric cholestasis.